a study on the correlations between seismotectonic -value ... · regions. in addition, a few new...

AbstractThe Western Anatolian region is one of the most seismically active sectors of Turkey where small-magnitude events with consi-

derable clustering occur frequently. The spatial and temporal behaviors of seismic activity as quantified by the fractal dimension Dc-value, seismotectonic b-value and seismic quiescence Z-value are investigated for the Western Anatolian region. For this pur-pose, a statistical relationship is developed between b and Dc-values, and regional and temporal changes of these parameters are analyzed in order to reveal the potential of future earthquakes in the study region. By using standard deviate Z-value, current seis-micity rate changes are estimated in the beginning of 2014.

The Western Anatolian region is divided into 18 different seismogenic subregions. Orthogonal regression fit is preferred in or-der to estimate more up-to-date and reliable statistical correlation between two seismotectonic parameters. The relationship of Dc=2.74-0.29*b is computed with a significant negative correlation (r = -0.73) between b and Dc-values for the Western Anatolian earthquake distributions. There are clear fluctuations in the temporal changes of these two parameters and the same results are relatively obtained in previous studies in literature. Lower b-values smaller than 1.0 and also higher Dc-value are observed in the Burdur fault zone and Kütahya graben. In addition, the regions in which the lower b-values and the higher Z-values are estimated cover the Burdur fault zone, Aliağa-Dumlupınar faults and Bakırçay grabens. Consequently, the relationships between seismic b-value, fractal dimension Dc-value and seismic quiescence Z-value may provide important evidences to put forth the next earth-quake potential in Western Anatolia. In addition, monitoring of microseismic activity and behaviors of some other geophysical parameters should be analyzed in space and time and evaluated carefully.

__________________________________________________________

1. IntroductionA number of statistical models have been proposed to des-

cribe the physical behaviors of earthquakes by using scaling laws in seismology (e.g., Mandelbrot, 1982; Hirata, 1989; Ön-cel and Wilson, 2007; Öztürk, 2011; 2012; Roy et al., 2011). In order to evaluate the spatial and temporal behaviors of seis-mic activity, some important seismotectonic parameters can be used. The parameters used in the scope of this study can be given as (1) b-value which describes the power-law distri-bution of earthquakes, (2) Dc-value which implies the num-ber of objects greater than a specified size and has a power law dependence on the size, and (3) Z-value which is one of the statistical parameters frequently used for analyzing the seismicity rate changes.

Seismically active fault regions are complex natural systems and they exhibit scale-invariant or fractal correlation between earthquakes in space and time (Öncel et al., 1995). Fractal di-mension Dc-value describes the heterogeneity degree of seis-micity in active fault system and some geological, mechanical or structural variations in heterogeneity. Thus, the higher or-der fractal dimension is increasingly sensitive for the distribu-tion of magnitudes (Öztürk, 2011). Estimating of b-value re-fers a fractal correlation between frequency of earthquake and the seismic moment, energy, or fault length. The b-value

for a region does not reflect only the relative proportion of the number of strong and small earthquakes in the region, but is also related to the stress condition over the region. Con-sequently, b-value is one of the most widely used seismicity parameters describing the size scaling behavior of earthqua-kes. In various parts of the world, a large number of studies on seismic quiescence analysis have been made in order to detect to precursory anomalies of a specific region. Wiemer and Wyss (1994) defined the precursory quiescence hypothe-sis in the following way: “A statistically significant decrease of the seismicity rate that occurs in a restricted segment of a seis-mogenic zone. The rate decrease is terminated by a main shock and the quiescent volume covers all or a major part of the source volume.” So, particular space-time seismicity occurrences that include the seismic quiescence phenomenon can be related to the seismotectonic processes that lead to earthquakes.

Turkey is a seismically very active region and therefore nu-merous statistical and physical studies have been carried out in order to examine the seismicity characteristics for seismic hazard assessments in many parts of Turkey, especially in the Western Anatolian region (e.g., Polat et al., 2008; Öztürk et al., 2008; Sayıl and Osmanşahin, 2008; Öztürk, 2012). However, detailed studies that represent possible correlations between

__

Austrian Journal of Earth Sciences Volume 108/2Vienna 2015 DOI: 10.17738/ajes.2015.0019

A study on the correlations between seismotectonic b-value and Dc-value, and seismic quiescence Z-value in the Western Anatolian region of Turkey

Serkan ÖZTÜRKGümüşhane University, Department of Geophysics, TR-29100, Gümüşhane, Turkey;

[email protected]

Western Anatolian region; fractal dimension; b-value; seismic quiescenceKEYWORDS

172

172 - 184

fault distribution and seismicity are limited. It is well known that the Aegean extensional region is one of the most seis-mically active and rapidly prolongating areas of the Eastern Mediterranean region (Bozkurt, 2001). So, this region was struck by many strong and destructive earthquakes in the past. Therefore, some potential applications to spatial and temporal behaviors of fractal associations between seismo-tectonic arguments may provide important contributions for assessments of earthquake occurrences. Statistical scaling properties of seismicity patterns may have a potential at least to be sensitive short term predictors of major earthquakes.

Therefore, the principal aim of this study is to analyze the spatial and temporal behavior of seismicity in order to reveal the future earthquake potential in the Western Anatolian re-gion. For this purpose, spatial and temporal assessments of the size-scaling distributions for this high-risk region are made by using some parameters such as seismotectonic b-value, fractal dimension Dc-value, completeness magnitude Mc-value and precursory seismic quiescence Z-value. In the final-ly, an up-to-date and a reliable empirical relation between b-value and Dc-value are presented for the Western Anatolian region of Turkey.

2. Seismotectonic Structure and Zonation in the Wes-tern Anatolian Region

The Western Anatolian graben systems and Aegean arc are one of the most important tectonic structures and seismically active fault regions in the Western Anatolia. In the Eastern Me-diterranean, there is a convergence between Anatolian and African plates caused by subduction of Cyprus and Aegean arcs, where the African plate is subducting beneath the Ana-tolian Plate in north-northeast direction (Bozkurt, 2001). In the geodynamical evolution of the Aegean extension region, the Aegean arc has an important effect. The nature and struc-ture of this region vary along the Aegean arc system and the eastern part of this system shows a transform fault characte-ristic. In addition, the dextral strike slip faulting mechanism associated with the North Anatolian Fault continues across the northern Aegean, crosses northern and central mainland Greece as a broad shear and eventually links up with the Hel-lenic subduction zone. This fault zone is also characterized by several second order faults that splay into the Anatolian Plate (Bozkurt, 2001). In Eastern Anatolia, there is convergence between the Eurasian and Arabian plates and as a result of this movement the Anatolian plate moves to the west along the North Anatolian Fault Zone and the East Anatolian Fault Zone. The Anatolian Plate moves anti-clockwise with a mean velocity of 24 mm/yr and this process continues along the Aegean in the southwestern direction (McClusky et al., 2000). Oral et al., (1995) stated that continental extension rate is ap-proximately 30-40 mm/yr in the north-south direction.

As shown in Figure 1, the most important faults in the Wes-tern Anatolian region are Bakırçay, Kütahya, Gediz, Simav, Küçük Menderes and Büyük Menderes (east-west trending), Dinar, Alaşehir and Akşehir-Afyon (northwest-southeast tren-

__

_____

ding), Burdur, Acıgöl, Çivril, Sandıklı-Dombayova (northeast-southwest trending), Soma, Zeytindağ-Bergama (north-south trending). These active faults show mainly normal faulting me-chanism. The Western Anatolian extensional region has expe-rienced many strong and destructive earthquakes during his-torical and instrumental epochs. Some of these earthquakes are 18 November 1919 Soma (M=6.9), 31 March 1928 Torbalı (M=6.3), 22 September 1939 Dikili-Bergama (M=6.5), 6 Octo-ber 1942 Edremit Körfezi–Ayvacık (M=6.8), 16 July 1956 Söke– Balat (M=7.1), 28 March 1969 Alaşehir (M=6.5), 28 March 1970 Gediz (M=7.2), 11 October 1986 Çubukdağ (M=5.5) , 6 Novem-ber 1992 Seferihisar-İzmir (M=6.0), and 10 April 2003 Seferih-isar-İzmir (M=6.1) earthquakes.

The seismotectonic boundaries of the study area are upda-ted from Öztürk (2012). Turkey is divided into 55 different source zones in Öztürk (2012) by taking into consideration the zonation studies by Erdik (1999) and Bayrak et al., (2009). Erdik et al., (1999) defined 37 source zones using all the avai-lable published data. In contrast, Bayrak et al., (2009) divided Turkey into 24 different source regions considering the diffe-rent zonation studies given above for seismic hazard mode-ling in Turkey and solution of focal mechanism given by TU-BITAK Marmara Research Center (Scientific and Technological Research Council of Turkey) for the great earthquakes that occurred in Turkey between 1977 and 2002, and plotting the existing tectonic structure with the epicenter distribution of earthquakes. So, some parts of these seismogenic zones for the Western Anatolian region are considered as the study

_________________________

Serkan ÖZTÜRK

Figure 1: Major tectonic structures in the Western Anatolian region. Principle faults were modified from Şaroğlu et al., (1992) and Bozkurt (2001). Names of the faults: EF: Etili Fault, YGSF: Yenice-Gönen and Sarıköy Faults, BG: Bakırçay Graben, SmG: Soma Graben, SG: Simav Graben, KG: Kütahya Graben, EİDKF: Eskişehir, İnönü-Dodurga and Kaymaz Faults, ZBF: Zeytindağ-Bergama Faults, ADF: Aliağa and Dum-lupınar Faults, GG: Gediz Graben, AG: Alaşehir Graben, AAG: Akşehir-Afyon Graben, BTKF: Beyşehir, Tatarlı and Kumdanlı Faults, KM: Küçük Menderes, BM: Büyük Menderes, BFZ: Burdur Fault Zone, SDG: San-dıklı and Dombayova Grabens, MRR: Muğla and Rhodes Region, GÇF: Gölhisar and Çameli Faults, MKFF: Marmaris, Köyceğiz and Fethiye Faults, ADÇG: Acıgöl, Dinar and Çivril Faults_____________________

173


regions. In addition, a few new smaller zones are added in order to compare the different tectonic structures in details in the same regions. Consequently, the Aegean extensional region of Turkey limited by the coordinates 25°N and 31°N in latitude and 35°E and 40°E in longitude is divided into 18 new seismotectonic subregions. Detailed tectonic structures are modified from Şaroğlu et al., (1992) and Bozkurt (2001). Ma-jor tectonics and selected 18 new seismic source zones in the Western Anatolian region are shown in Figures 1 and 2, res-pectively.

3. Earthquake databaseThe local magnitude, M , is one of the various types of mag-L

nitudes to calculate the quantitative size of the earthquakes. Magnitudes of earthquakes are actually based on the ampli-tude of ground motion displacement as measured by a stand-ard seismograph. The Richter magnitude is one of best known among them and it was originally defined for local events for the southern California. M is often used by local seismic net-L

work and given a formula by Richter (1935; 1938) as in the following:

(1)

Where A(Δ) is the maximum earthquake amplitude in mil-limeters recorded on the standard Wood-Anderson torsion seismograph which has a static magnification of 2800, a dam-ping factor 0.8, and a natural period of 0.8 second at an epi-central distance Δ in kilometers. A describes the loss of ener-0

gy with respect to distance such as geometrical spreading, anelastic attenuation due to the local geology. C is an empi-rical correction for the particular station or instrument used and it is strongly depended on the local geology. However, this magnitude scale has also been used in the other seismic network in the world to determine the local magnitude be-cause it is presumed that the Richter’s southern California

Log A has the similar properties in the crust and mantle.10 0

A few earthquake databases are available for Turkey from both national and international catalogs. The database used in this study is taken from Öztürk (2009) for the time period between 1970 and 2006 (all details for the relationships of different magnitude types can also be found in Bayrak et al., 2009). He used some empirical relationships in order to pre-pare a complete and homogeneous earthquake catalog, and prepared an instrumental database for duration magnitude M including 73,530 earthquakes between 1970 and 2006. D

The Bogazici University, Kandilli Observatory and Research Institute (KOERI) catalog is also used for the data from 2006 to 2014. KOERI generally provides the type of M for all earth-D

quakes, especially after 2000. However, KOERI usually gives local magnitude M for missing MD in recent years. In such a L

case that M is unknown in KOERI catalog from 2006 to 2014; D

calculations of unknown M are made by using M -M rela-D D L

tionships (Table 1) given in Öztürk (2009). 76,735 earthquakes are finally obtained in and around Turkey from 2006 to 2014. Thus, 150,265 earthquakes are obtained in totally for all re-gions of Turkey between 1970 and 2014.

After the selection of the Western Anatolian region as the study area, the earthquake catalog for this region is prepared. In this stage, the earthquakes from 1970 to 2014 are selected for the studied region from whole catalog. The type of mag-nitude used in this study is M and the study catalog is com-D

plete for the whole study period and for all magnitude levels. Thus, the final data catalog is prepared in the time period from 6 January 1970 to 31 December 2013, the time interval of about 43.98 years. Finally, there are in total 79,881 events with magnitudes larger than or equal 1.0 in this coordinates during 1970- 2014. Epicenter locations of all earthquakes for M ≥1.0 in the Western Anatolian region for this time inter-val D

are also shown in Figure 2.

4. Brief Descriptions of the Statistical MethodsIn the scope of this study, a few statistical arguments such

as seismotectonic parameter b-value of Gutenberg-Richter (G-R) relation, fractal dimension Dc-value and seismic quie-scence Z-values are analyzed as the spatial and temporal va-riations of seismicity in the Western Anatolian Region.

4.1 Gutenberg-Richter relation (b-value) and magni-tude completeness (Mc-value)

The relation between magnitude and frequency of occur-rence of earthquakes was described by Gutenberg-Richter (1944) as in the following equation

(2)

where N(M) is the expected number of earthquakes with mag-nitudes equal to or greater than M, b-value describes the slope of the magnitude-frequency distribution, and a-value is pro-portional to the activity rate of seismicity. a-value changes from region to region. These variations depend on the obser-

_________________

_____

______________________

Figure 2: Seismotectonic zones and the epicenters of 79,881 earth-quakes with M ≥1.0 between 1970 and 2014D __________________

M =Log A(Δ)-Log A (Δ)+CL 10 10 0

log N(M)=a-bM10

174

Serkan ÖZTÜRK

vation period, length of the stu-dy area and also size of earth-quakes. Utsu (1971) stated that parameter b varies roughly be-tween 0.3 and 2.0, depending on the different regions. How-ever, the changes in b-value can be caused by many factors such as the number of small and great events, geological complexity and degree of heterogeneity of cracked medium, strain and stress condition in the region (e.g., Schorlemmer et al., 2005). The regional scale estimates of b-value are approximately equal to 1 on average (Frohlich and Davis, 1993).

Completeness magnitude Mc is an important parameter in many seismicity studies, especi-ally in investigation of magni-tude-frequency relation. The power law distribution of G-R against magnitude is used in order to estimate Mc-value. The variation in Mc-value is calcu-lated using a moving time win-dow approach (Wiemer and

_______________

Table 1: M and M relations for 24 different seismotectonic zones of Turkey. The values in the parenthe-D L

ses show the uncertainties (from Öztürk, 2009)

RegionNumber

123456789

101112131415161718192021222324

EarthquakesNumber

201411244

26142

11238146122970156712186222171121

Empirical relations

MD = 0.881(±0.138)*ML+0.596(±0.286)MD = 0.919(±0.023)*ML+0.292(±0.048)MD = 0.991(±0.080)*ML+0.033(±0.158)MD = 0.768(±0.114)*ML+1.004(±0.239)

-MD = 0.816(±0.068)*ML+0.825(±0.147)MD = 0.812(±0.112)*ML+0.726(±0.234)

-MD = 0.432(±0.339)*ML+2.293(±0.675)MD = 0.843(±0.066)*ML+0.580(±0.137)MD = 0.818(±0.036)*ML+0.586(±0.075)MD = 1.277(±0.209)*ML-1.372(±0.434)MD = 1.113(±0.389)*ML-0.555(±0.768)

MD = 0.956(±0.057)*ML+0.103(±0.114)MD = 0.934(±0.029)*ML+0.163(±0.062)MD = 0.446(±0.146)*ML+1.900(±0.291)MD = 0.748(±0.043)*ML+0.869(±0.089)MD = 0.886(±0.044)*ML+0.349(±0.087)MD = 0.901(±0.049)*ML+0.268(±0.100)MD = 0.939(±0.068)*ML+0.091(±0.138)MD = 0.876(±0.069)*ML+0.450(±0.139)MD = 0.873(±0.043)*ML+0.467(±0.089)MD = 1.229(±0.691)*ML-0.707(±1.382)MD = 0.743(±0.099)*ML+1.211(±0.222)

CorrelationCoefficient

0.8200.9960.9660.808

-0.9200.889

-0.3590.9350.9290.6690.6360.9520.9670.6190.9030.9850.9740.8670.9390.9800.4730.851

Dc=lim[logC(r)/logr]r0

C(r)=2N /N(N-1)R<r

Wyss, 2000). If the completeness magnitude changes syste-matically as a function of time and space, temporal variations of Mc-value can cause potential wrong value of seismicity pa-rameters, primarily in b-value. Thus, Mc analysis of the cata-log used in this study is an im-portant process since a part of this study uses Mc-value in the estimation of b-value.

4.2 Fractal dimension (Dc-value)To evaluate the size scaling attributes and clustering pro-

perties of seismotectonic arguments, fractal analysis is often used. Spatial and temporal patterns of earthquake occurrence are demonstrated to be fractal using the two-point correla-tion dimension Dc. Correlation dimension Dc and the correla-tion sum C(r) was defined by Grassberger and Procaccia (1983) as in the following equations:

(3)

(4)

where C(r) is the correlation function, r is the distance between two epicenters and N is the number of events pairs separated by a distance R<r. If the epicenter distribution has a fractal structure, following relation is obtained:

(5)

______

__________________________

_________________

where Dc is a fractal dimension, more strictly, the correlation dimension. The distance r (in degrees) between two earth-quakes is calculated from:

(6)

where (Θ ,Φ ) and (Θ ,Φ ) are the latitudes and longitudes of i i j jth ththe i and j events, respectively (Hirata, 1989). By plotting

C(r) against r on a double logarithmic coordinate, fractal di-mension Dc is practically obtained from the slop of the graph.

Fractal dimension characterized the nature of spatial and temporal properties of the earthquakes. Dc is calculated to evaluate the possible unbroken sites and seismic gaps, which may be broken in future (Kagan, 2007). In other words, the fluctuations in fractal properties principally depend on the complexity or quantitative measure of the degree of hetero-geneity of seismic activity. Higher Dc-values associated with lower b-values are the dominant structural feature in the areas of increased complexity in the active fault system and it may be caused due to clusters. Thus, this property may be an indication of stress changes on fault planes of smaller sur-face area (Öncel and Wilson, 2002; Polat et al., 2008).

4.3 Declustering of earthquake catalog and defini-tion of the seismic quiescence method (Z-value)

The algorithm of cluster analysis “declusters” or decomposes

_____________________________

______

___DcC(r)~r

-1r=cos (cosΘ cosΘ+sinΘ sinΘ cos(Φ -Φ ))i j i j i j

175


an earthquake catalog into main and secondary events (Ara-basz and Hill 1996). This process removes all dependent events from each cluster and substitutes them with a unique event. Eliminating the dependent events from the catalog is necessary for a quantitative seismic quiescence analysis. In this study, ZMAP software introduced by Wiemer (2001) is used to decluster the earthquake catalog based on the algo-rithm developed by Reasenberg (1985) and this declustered catalogue is used for making seismic quiescence analysis. In recent years, there were a considerable number of investiga-tions of seismic quiescence phenomena by using ZMAP soft-ware (e.g., Wyss et al. 2004, Polat et al. 2008; Öztürk, 2011; 2013; Öztürk and Bayrak, 2012).

There are 79,881 events with magnitudes equal to or larger than 1.0 in the catalog. Average Mc-value for whole study re-gion from 1970 to 2014 was calculated as 2.8 and the number of events with magnitude M <2.8 are obtained as 35,043. All D

events with magnitude M <2.8 are subtracted from the cata-D

log and thus, the number of earthquakes exceeding this mag-nitude threshold is found as 44,838. Using the declustering algorithm, 8973 (about 20%) earthquakes are subtracted and about 55% of all earthquakes in totally were taken away from the catalog. Thus, the number of events for Z-test in the Wes-tern Anatolian region was reduced to 35,865 with M ≥2.8.D

There are many techniques describing the seismicity rate change and most of them use the phenomenon of seismic quiescence and the most frequently used is the standard de-viate Z-test. ZMAP technique is used in order to image the re-gions displaying seismic quiescence (for details see Wiemer and Wyss, 1994). Z-test generates the long term average, LTA , (t)

function for the statistical evaluation of the confidence level in units of standard deviations:

(7)

where R is the mean seismicity rate in the foreground window, 2

R is the average number of earthquakes in the whole back-1

ground period, S and N are the standard deviations and the number of samples, within and outside the window. The Z-value computed as a function of time, letting the foreground win-

________________________

___

_________________________

dow slide along the time duration of catalogue, is called LTA.

5. Results of Spatial and Temporal Analyses and Dis-cussions

In the scope of this study, the Western Anatolian region is divided into 18 different seismotectonic zones to make a de-tailed statistical analysis among two seismotectonic parame-ters, b and Dc-values. The calculation of b and Dc-values for 18 regions is carried out using ZMAP software (Wiemer, 2001). The maximum likelihood method is used in order to calculate the b-values, because it yields a more robust estimate than the least-square regression method (Aki, 1965). The Dc-values for all parts of the Western Anatolian region are obtained with 95% confidence limits by linear regression.

Mc usually shows a non-stable value in time and has a great importance for many seismic studies. It is very important to use the maximum number of earthquakes for high quality results in seismicity analyses. In this study, the change of Mc-value as a function of time is estimated by using a mowing window approach with maximum curvature method (Woess-ner and Wiemer, 2005). The whole earthquake catalog contai-ning all 79,881 events with M ≥1.0 is included in the estima-D

tion of Mc-value and it is plotted with its standard deviation for samples of 250 events/windows. Figure 3 shows the chan-ges in Mc-value with time. Mc-value is rather large and varies from 3.5 to 4.5 until 1975 whereas it changes from about 3.5 to about 2.5 between 1975 and 1990. Mc has a value between 2.7 and 3.0 from 1990 to 2010. Then, Mc-value is smaller than 2.7 after 2010. It can be easily seen that Mc-value changes between 3.0 and 2.5 after 2000. As a result, Mc-value for the Western Anatolian region varies between 2.5 and 3.0 and this value is compatible with the results of Polat et al., (2008).

Figure 4 shows the cumulative number of events versus time for the original catalog including all dependent 79,881 events with M ≥1.0 and for the declustered catalog including D

35,865 independents events with M ≥2.8 between 1970 and D

2014. As shown in Figure 4, there is no significant seismic acti-vity from 1970 to 1976 and a little seismicity change between 1976 and 1995. On the contrary, there are great seismic chan-ges after 1995. The cumulative earthquake number of declus-

_______________

___

Figure 3: Variations of magnitude completeness Mc-value from 1970 to 2014 in the Western Anatolian region. Standard deviation (δMc) of the completeness (dashed lines) is also given. Mc is plotted for over-lapping samples, each containing 250 events

2 2 1/2Z=(R -R )/(S /N +S /N )1 2 1 1 2 2

Figure 4: Cumulative number plots of the earthquakes versus time for all events with M ≥1.0 and declustered events with M ≥2.8 in the D D

Western Anatolian region

176

Serkan ÖZTÜRK

tered catalog with M ≥2.8 as a function of time for the Wes-D

tern Anatolian region has a smoother slope when compared to the original catalog. Thus, it is clear from Figure 4 that de-clustering algorithm has removed dependent events from the original catalog and after processing the declustering algorithm, a more homogeneous, reliable and robust earth-quake catalog has been obtained.

The resulting values of two seismotectonic parameters for all tectonic subregions are shown in Table 2: b-value changes between 0.97 and 1.95; b-values smaller than 1.0 are found in regions 7 and 13, including Burdur Fault zone and Kütahya graben. Moderate values around 1.0 are calculated along the Aliağa and Dumlupınar Faults, Aegean Arc and region 18 (wes-tern sea part of Bakırçay graben). The highest b-value is cal-culated as 1.93 in region 16, including Zeytindağ-Bergama faults. Other high values above 1.5 are estimated in regions 5, 15 and 17 - these regions are related to the Muğla and Rhodes region, Soma and Bakırçay grabens. The b-values obtained for the rests of the subregions range from 1.2 to 1.5. Regional distribution of b-value is also plotted at every node of the 0.05º grid (Fig. 5). As for changing Mc-values, variations in b-values are estimated by using a moving window approach. Declustered earthquake catalog with M ≥2.8 is included in D

the estimation of b-value with samples of 500 events/win-dows. As stated in Frohlich and Davis (1993), average b-value estimation is approximately equal to 1.0. The Western Anato-lia is an extensional zone and has relatively large b-values since the stress is more easily reduced caused by great num-ber of small earthquakes and this situation can be explained by large heterogeneity (Polat et al., 2008), low stress distribu-tion due to high heat flow (İlkışık, 1995). According to Scholz (1968), lower b-values indicate higher stress release. Conse-quently, low b-values in the Burdur Fault zone and Kütahya graben may be an indication of low degree of heterogeneity, high-strain due to the subduction tectonics and stress to build up over time and to be released by earthquakes that are less frequent but great in magnitude (Öncel and Wilson (2002). In the regions where larger b-values are estimated, however, this situation may be considered as an indication of low stress re-lease by a large number of small earthquakes and thus geo-logical complexity is very high (Lopez Casado et al., 1995). As a general result, estimated b-values through the maximum likelihood approach for the G-R method seem to have a good relation to the tectonics and seismic activity. Thus, special interest should be given to the regions where low b-values are observed.

Fractal dimension Dc-value varies from 2.11 to 2.51 for all 18 seismotectonic subregions as seen in Table 2. Dc-values are below 2.3 in regions 5, 12, 14 and 16. These regions cover Muğla and Rhodes, Gediz, Alaşehir and Simav grabens, and Zeytindağ-Bergama faults. Dc-values changing between 2.3 and 2.4 are obtained in regions 1, 2, 3, 8, 9, 10, 15 and 17 which includes Sandıklı and Domboyova grabens, Acıgöl, Dinar, Çivril, Gölhisar ve Çameli faults, Büyük and Küçük Menderes grabens, western part of Aliağa and Dumlupınar faults, Soma and Ba-

______________________

kırçay grabens. In all the remaining regions, Dc-values fluc-tuate between 2.4 and 2.5 (regions 4, 6, 7, 11, 13, 18). These zones are related to Marmaris, Köyceğiz and Fethiye faults, Aegean Arc, Burdur fault zone, Aliağa and Dumlupınar faults. Estimated Dc-values in the Western Anatolia are generally higher than 2.2. As in b-value map (Fig. 5), the regional distri-bution of Dc-values is also plotted at every node of 0.05º grid (see Fig. 6). These results reveal that the earthquake distribu-tion becomes less clustered (higher Dc-value) since the proba-bility of strong earthquakes becomes smaller (larger b-value) and this indicate that, in general, there is greater fracture

Figure 5: Spatial distribution of seismic b-value for the Western Ana-tolian region. White dots show the declustered earthquakes with M ≥2.8D

Figure 6: Spatial distribution of fractal dimension Dc-value for the Western Anatolian region. White dots show the declustered earth-quakes with M ≥2.8D

177


toughness in all parts of the Western Anatolia. Barton et al., (1999) stated that the faults where earthquakes are caused by failure of isolated, small asperities and occurred in clusters, are related to the higher b-value and lower Dc-value. The hig-her order fractal dimension (especially greater than 2.3) is in-creasingly sensitive to heterogeneity in the distribution of magnitudes. This suggests that seismicity is more clustered at larger scales (or in smaller areas) in these regions. However, there are only two regions whose b-values are smaller than 1.0 and in addition, higher Dc-values larger than 2.4 are esti-mated in these zones. As stated above, these zones are rela-ted to the Burdur fault zone (region 7) and Kütahya graben (region 13). Öncel and Wilson (2002) stated that, in the regions of increased complexity in the active fault system (higher Dc-value) associated with lower b-value, the stress release occurs on fault planes of smaller surface area. As a result, it is reaso-nable to assume that the higher Dc values (≥2.4) and lower b-values (≤1.0) are the dominant structural feature in the study area and may arise due to clusters since the uniform distribu-tion of events decreases with an increase in the clustering of earthquakes.

To determine a reliable and suitable statistical relationship between two seismotectonic parameters b and Dc-values for the Western Anatolia, the orthogonal regression (Carrol and Ruppert, 1996) is used since the standard least square method is based on the assumption that horizontal axis values are estimated without error. Figure 7 shows the relationship of orthogonal regression fit between b and Dc-values. Also, Dc-value versus b-value for orthogonal regression method with fit curve, corresponding equation and, 95% confidence inter-val are given (Fig. 7). Negative correlation coefficient (r) is cal-culated as 0.73 and it can be accepted as a strong enough value. There are 13 earthquakes within the confidence limit of the regression. Linear regression fit is used and following relationship is obtained:

(8)

Many examples of previous studies on the correlation be-tween Dc and b-values can be found for different parts of the

______________________________

world (e.g., Aki, 1981; Hirata, 1989; Öncel et al., 2001; Roy, et al., 2011) and Turkey (e.g., Öncel et al., 1995; 1996; Öncel and Wilson, 2002; 2007; Öztürk, 2012). Since Aki (1981) suggested a simple relationship between b and Dc-values with a positive correlation D=3b/c (where c is a constant determined from the slope of the log moment versus the magnitude relation, c is normally taken as 1.5), both positive (e.g., Öncel and Wil-son, 2004; Roy et al., 2011) and negative (e.g., Hirata, 1989, Öncel et al., 1995; 1996) correlations between these two seis-motectonic parameters have been reported for different parts of the world and Turkey.

Hirata (1989) pointed out that Aki's fractal dimension cor-responds to the capacity dimension and may be compared with the correlation dimension. Hirata’s (1989) result do not support Aki’s speculation that D=3b/c, on the contrary, there is a negative correlation as Dc=2.3-0.73*b (with r=-0.77) be-tween b and fractal dimension of epicenters in the Tohoku region of Japan. Similarly, a study of seismicity in the North Anatolian Fault Zone (NAFZ), Turkey, revealed a long- term negative correlation between b and Dc (Öncel et al., 1995). The b-value is found to be weakly negatively correlated with fractal dimension as Dc=2.74-1.52*b (with r=-0.56) for the NAFZ (including the northern Aegean sea) in Öncel et al., (1995). Also, Öncel et al., (1996) investigated the nature of temporal variations in the statistical properties of seismicity associated with the NAFZ during the instrumental period 1900-1992 and observed a strong negative correlation (r=-0.85) between Dc and b-values as Dc=2.32-1.09*b. In contrast, Öncel and Wilson (2002) made an analysis that is restricted to the NAFZ. They found a weak positive correlation (r=0.48) between variations in b and Dc in the western NAFZ. Analysis presented in Öncel and Wilson (2004) reveals a strong positive correlation (r=0.81) between Dc and b-values along the NAFZ during the 1981 to 1998 time period preceding the 1999 Izmit earthquake. Öncel and Wilson, (2007) observed a strong positive correlations between Dc and b-value during the 1992-1994.4 (r=0.84) and 1996.6-1998.2 (r=0.94) and negative correlation (r=-0.71) ex-tending from approximately mid-1994 to mid-1996 in the northwestern Turkey between 40.5° to 41° north latitude, and 29° and 31° east longitude. Consequently, estimated relation, correlation coefficient and the number of events in the confi-dence limit from orthogonal regression provides a reliable

Figure 7: Orthogonal regression fit, confidence interval, correspon-ding equation and correlation coefficient for the relationship b-value and Dc-value for the Western Anatolian region. There are 13 events in the confidence interval___________________________________

Figure 8: Magnitude distribution as a function of time for earthquakes occurred in the Western Anatolian region from 1970 to 2014________

Dc=2.74-0.29*b, (r=0.73)

178

Serkan ÖZTÜRK

assessment for the seismotectonics of the Western Anatolian region. Also, this equation is basically similar to the results presented in the literature.

The seismotectonic parameters b and Dc-values as well as the numbers of earthquakes, minimum (M ) and maximum min

(M ) magnitudes, Mc-values and, a-values for every one year max

are given in Table 3. In order to make an assessment on the seismic activity with time, temporal distribution of all M ≥1.0 D

earthquakes from 1970 to 2014 is plotted as seen in Figure 8. As shown in Table 3, there are a few strong earthquakes lar-ger than 5.5 from 1970 to 2000. But further on, significant fluctuations in seismic activity are recorded after year 2000. Magnitude-time histogram indicates great seismic changes in the number of strong events especially from 2005. The seismic activity related to the clustering properties is clearly observable and this situation may correspond to a main event in the region. In addition, temporal clustering characteristic of the seismic activity related to the major events is strong enough for many earthquakes in which occurred 1975, 1981, 1997, 2002, 2003, 2005, 2008, 2011, 2012 and 2013. Temporal changes in b and Dc-values versus time are plotted in Figure 9. These variations as a function of time are estimated to in-vestigate the possible temporal changes during the time pe-riod 1970 and 2014. While Dc-values show a strong increa-sing trend before some certain years, b-values show a strong decreasing trend in the same years. Also, these fluctuations can be clearly seen in Table 3 and from arrows on Figure 9. For example, Dc-value is increasing while b-value is decrea-sing from 1974 to 1975 and, a larger earthquake than that of 1974 occurred in 1975. Such kind of similar changes are also estimated between 1978 and 1979, 2000 and 2001, 2004 and 2005, 2006 and 2007. However, such kinds of changes are not observed from 1997 to1998 and from 2011 to 2012. Many fac-

tors can cause perturbations of these parameters as mentio-ned above. In active fault system, stress release occurs on fault planes of smaller surface area since the higher Dc-values are associated with lower b-value. The higher order fractal di-mension is increasingly sensitive to heterogeneity in the dis-tribution of magnitudes. This suggests that seismicity is more clustered at larger scales (or in smaller areas) in the West Ana-tolian region.

In order to map the spatial distribution of the Z-value, the study area is firstly divided into rectangular cells spacing 0.05º in latitude and longitude. The nearest earthquakes, N, at each node are taken as 50 events. Then, the changes in seismic ac-tivity rate are searched within a maximum radius changes by a moving time window T (or iwl), stepping forward through W

the time series by a sampling interval as described by Wiemer and Wyss (1994). In order to have a continuous and dense co-verage in time, population of the events is binned into many binning spans of 28 days for each grid point. T =5.5 years are W

used as the window length because the quiescence areas are better visible for a window of 5.5 years.

Regional variations of Z-value with T =5.5 years is plotted W

in Figure 10. The length of T for Z-value map is calculated by W

adding T -value to the time chosen as the beginning of W

the time cut as indicated in top of the Figure 10. Thus, Z-value map is plotted for the beginning of 2014. Some clear seismic quiescence regions in the Western Anatolia are defined. These anomalies regions are pointed out around and in the north-east of Simav fault (region 14), between Soma (region 15) and Bakırçay grabens (region 17), around Aliağa-Dumlupınar faults (region 11) and Alaşehir graben (region 12), in the western end of Gediz graben (region 12), in the junction of Gediz, Kü-çük and Büyük Menderes grabens (region 8), between Gölhi-sar-Çameli faults and Muğla-Rhodes region (region 3), around

__________________

Table 2: Seismotectonic parameters b and Dc-values with their standard deviations as well as the number of earthquakes and completeness mag-nitudes for all tectonic subregions in the Western Anatolian region

Region123456789

101112131415161718

Earthquake Numbers255020437532262523471751203942362650105347772402768483519123679022371574

Mc value2.93.03.02.83.53.52.82.92.93.13.12.82.62.82.72.73.03.3

b-value1.29±0.051.47±0.091.49±0.051.24±0.061.63±0.091.19±0.070.98±0.051.28±0.071.37±0.061.34±0.061.18±0.121.25±0.050.97±0.151.24±0.051.63±0.051.95±0.061.63±0.071.19±0.04

Dc-value2.36±0.032.31±0.032.31±0.052.51±0.032.22±0.022.43±0.022.45±0.032.34±0.032.36±0.032.36±0.032.44±0.022.23±0.012.41±0.032.27±0.062.37±0.022.11±0.032.39±0.022.44±0.04

Tectonic Environments

Gölhisar and Çameli Faults,Marmaris, Köyceğiz, Fethiye Faults

and Muğla and Rhodes RegionAegean Arc

Burdur Fault ZoneBüyük and Küçük Menderes

Grabens

Gediz Graben and Alaşehir GrabenKütahya Graben

Simav GrabenSoma Graben

Zeytindağ-Bergama Faults

Bakırçay Graben

Sandıklı and Dombayova Grabens and Acıgöl, Dinar and Çivril Faults

Aliağa and Dumlupınar Faults

179

Year

19701971197219731974197519761977197819791980198119821983198419851986198719881989199019911992199319941995199619971998199920002001200220032004200520062007200820092010201120122013

EarthquakeNumbers

238913528

221166

1241840562

1016616697787599

1062847

1310597669

1312161011641350237525853295257324621389139311811125134413572526418813361449241829453856616089267940

Mmin

4.04.03.93.92.82.72.62.72.32.22.01.81.81.92.02.12.02.12.12.12.01.72.22.32.32.32.32.22.42.22.42.22.22.12.32.22.32.32.32.02.11.61.01.0

Mmax

5.55.35.05.05.15.65.35.34.95.55.26.04.85.44.94.95.44.74.95.05.25.45.45.05.15.24.95.55.35.24.95.35.95.65.35.95.25.36.45.25.06.75.75.9

Mc-value

4.44.24.04.23.43.53.53.53.43.43.03.03.23.02.82.93.02.82.53.02.93.03.02.82.82.82.72.72.62.83.02.82.82.82.82.82.93.02.82.72.72.72.42.4

a-value

9.317.9813.27.115.595.677.865.867.356.705.325.265.355.104.905.546.444.324.925.225.185.096.207.717.006.937.197.276.396.407.366.196.406.826.936.877.106.716.816.497.107.286.245.57

b-value

1.65±0.101.57±0.041.09±0.121.40±0.301.54±0.091.21±0.071.48±0.061.38±0.081.66±0.111.20±0.041.19±0.071.17±0.051.24±0.041.19±0.091.14±0.101.09±0.061.18±0.040.77±0.100.95±0.081.39±0.111.16±0.101.42±0.131.51±0.091.49±0.081.36±0.031.45±0.051.43±0.031.49±0.041.27±0.031.48±0.061.50±0.051.17±0.041.20±0.031.07±0.131.30±0.031.21±0.021.63±0.071.26±0.041.45±0.091.35±0.111.38±0.021.39±0.021.15±0.111.48±0.08

Dc-value

2.54±0.032.46±0.031.50±0.011.50±0.011.55±0.011.64±0.031.77±0.021.82±0.021.97±0.022.03±0.031.92±0.022.13±0.072.01±0.022.00±0.031.92±0.041.79±0.021.96±0.041.52±0.031.17±0.031.54±0.031.39±0.031.53±0.032.54±0.032.51±0.032.23±0.032.23±0.032.22±0.032.15±0.022.37±0.052.24±0.042.36±0.032.58±0.042.61±0.022.31±0.052.54±0.042.72±0.042.14±0.032.29±0.062.43±0.042.31±0.052.20±0.052.20±0.052.31±0.052.16±0.05

Table 3: b and Dc-values and some other seismic parameters such as the number of earthquakes, mini-mum magnitudes, maximum magnitudes and a-value as a function time for the Western Anatolian region


Acıgöl, Dinar, Çivril faults (region 2), in the southeastern part of the Zeytindağ-Bergama faults (junction of regions 11 and 12), between Sandıklı-Domboyova grabens (region 1) and Burdur fault zone (region 7).

In addition to above-mentioned quiescence areas, there are a few anomalies in the western part of the study area. One of

___________________________

them is found centered at the coordinates 39°N in longitude and 25°E in latitude, 38°N in longitude and 25°E in latitude and nearly 37°N in longitude and 26°E in latitude. These areas are in the sea part of the Western Anatolian region, which has no fault system (regions 10 and 18). These values can be inter-preted as the artificial results from contouring or interpolations

since there are fewer earthqua-kes in these regions. As stated in Joswig (2001), characterizing the null hypothesis should be made before the interpretation of the seismic quiescence maps. This means the fraction of suc-cess for earthquake predictions and it can be achieved by pure chance. Such kind of quiescen-ce maps do not issue any alert but should help to relate quie-scence spots to pending earth-quakes. So, the null hypothesis describes how many quiescence anomalies would precede a real event, even if the earthquake distribution is completely ran-dom. The small scale quiescence anomalies in some regions can be interpreted as false alarms exceeding in significance the precursors. This randomness could be derived from the given catalogue by arbitrarily altering the event times, but keeping their locations for the spatial clustering (Joswig, 2001). Thus, such kind of heterogeneous re-porting as a function of time can generate false alarms and im-pede reliable measurement of natural seismicity rate changes.

Polat et al., (2008) made an earthquake hazard assessment for the Aegean extension region of Turkey by using fractal beha-vior, Gutenberg-Richter b-value and seismic quiescence Z-value. They found some anomalous regions including Çandarlı Bay and Bergama-Zeytindağ fault zone, İzmir and Orhanlı faults zone and, Buldan and surroun-ding regions. They suggested that the sites of larger Z-values and smaller b-values can be con-sidered to be the most likely regions for future earthquakes.

180

Serkan ÖZTÜRK

This evidence can be explained with most promising environ-ment where decrease in b-value is found with an increase in mean stress (Westerhaus et al., 2002). In this study, both the highest b-value and Z-value are observed in some regions in-cluding Simav and Alaşehir grabens, Gediz, Küçük Menderes and Büyük Menderes grabens, Gölhisar-Çameli faults and Muğla-Rhodes region, Acıgöl, Dinar, Çivril faults and, Sandıklı-Domboyova grabens. However, the lower b-values and the higher Z-values are observed in some regions including Bur-dur fault zone, Aliağa-Dumlupınar faults and Bakırçay grabens. Also, the findings and anomaly regions for b and Z-values in this study are more up-to-date and quite similar with those of Polat et al., (2008).

An investigation of the seismicity of Western Anatolia was made by Sayıl and Osmanşahin (2008), computing the para-meters of Gutenberg-Richter b-value, seismic risk and recur-rence period. According to their seismic risk estimations, the highest-earthquake occurrence probability of Ms≥7.0 in the next 100 years is 80.6±0.20% for their subregion 9 (around Bodrum-İstanköy) and 77.8±0.17% for their subregion 1 (in-cluding Balıkesir). Recurrence times for the earthquakes with the same magnitude have been found as 61 and 67 years in these subregions by Sayıl and Osmanşahin (2008). Their re-gions 1 and 9 cover Bakırçay graben (regions 17 and 18) and Muğla-Rhodes region (region 5) in this study. Although no anomaly is observed in Muğla and Rhodes region, the results in the regions including Bakırçay graben are similar with those of Sayıl and Osmanşahin (2008).

Öztürk et al. (2008) estimated the earthquake hazard para-meters for different regions in and around Turkey using Gum-bel’s first asymptotic distribution. They estimated the mean return periods, the most probable maximum magnitude and the probability of an earthquake occurrence for a given mag-nitude in a period of 10, 25, 50, and 100 years for 24 regions in and around Turkey. Their regions 13 (including Burdur fault zone), 15 (covering Gediz graben, Aliağa-Dumlupınar faults) and 17 (Kütahya-Simav and Zeytindağ-Bergama faults) are up-dated in this study in order to compare such tectonic zones in smaller scale and detail. The results from Öztürk et al. (2008) show that the mean return period for Ms≥5.5 in their region 13 covering the Burdur fault zone (region 7 in this study) is

equal to 24.55±0.28 years. For their region 15, including the Aliağa and Dumlupınar faults (regions 10 and 11 in this stu-dy), the value of the mean return period for Ms≥6.0 is calcula-ted as 23.44±0.06 years. In their region 17 covering Kütahya, Simav, Soma and Bakırçay grabens and Zeytindağ-Bergama faults (regions 13, 14, 15, 16, 17 and 18 in this study), the mean return period value for the earthquakes with Ms≥6.0 is computed as 18.62±1.29 years. According the results from Öztürk et al. (2008), these short mean return periods for the earthquakes between 5.5 and 6.0 in mentioned regions, es-pecially the Burdur fault zone, Bakırçay graben and Aliağa-Dumlıpınar faults, are important and in high risk. Thus, the next earthquake of such magnitude in these regions can be expected between 2017 and 2022.

Thus, spatial and temporal analysis of the current seismic behaviors of earthquakes may be the key to the future earth-quake potential in the Western Anatolian region. For this rea-son, size scaling parameters such as Dc-value, b-value, Z-value and their relations must be more carefully estimated in order to reveal the significant anomalies prior to a strong earthquake in the next future. There is a current quiescence in the begin-ning of 2014 and the other seismotectonic parameters cor-relate with this quiescence. Consequently, special attention must be given to these anomaly regions in Western Anatolia.

6. ConclusionsThis study focused on the spatial and temporal behaviors of

the seismic activity in the Western Anatolian region of Turkey. In this scope, statistical correlation between seismotectonic b-value and Dc-value is estimated and the current seismicity rate changes in the beginning of 2014 are mapped. Earth-quake catalogue, taken from KOERI, is homogeneous for du-ration magnitude, M and consists of 79,881 events with mag-D

______________________

Figure 9: Variations of the seismotectonic parameters b and Dc-val-ues as a function of time in the Western Anatolian region after 1970. Standard errors are also shown. Arrows indicate the beginning times of decreasing in b-value and increasing in Dc-value______________

Figure 10: Spatial distribution of seismic quiescence Z-value in the be-ginning of 2014 with T (iwl) equal to 5.5 years for the Western Anato-W

lian region. White dots show the declustered earthquakes with M ≥2.8D

181


nitudes between 1.0 and 6.7 from January 1, 1970 to January 1, 2014. Average Mc-value for whole study region from 1970 to 2014 is estimated as 2.8. Reasenberg’s algorithm is used to decluster the catalog and then seismic quiescence Z-value is calculated. The Western Anatolian region was divided into 18 different seismogenic subregions to make a comprehensive study. The maximum likelihood method is used to calculate the b-values and linear regression with 95% confidence limit is used to obtain Dc-values.

Significant changes in seismic activity started after 2000 and there was an increase in the number of earthquakes with mag-nitudes greater than 5.5. However, great seismic changes star-ted after 2005. Thus, seismic activity is more clustered at lar-ger scales (or in smaller areas) in the West Anatolian region. According to the results of regional and temporal changes in seismotectonic parameters b and Dc-values, there are clear fluctuations in the study area. So, there has been a strong earthquake potential in the Western Anatolia because of these recent seismic fluctuations. The Burdur fault zone and Kütahya graben in which b-values are smaller than 1.0 must be given a special caution, as higher Dc-values are associated with lower b-values. To estimate a more up-to-date and reliable statistical relation between two seismotectonic parameters b and Dc-values, orthogonal regression is preferred. The relationship of Dc=2.74-0.29*b is suggested with a strong enough negative correlation (r = -0.73) for the Western Anatolia earthquake distributions.

Regional variation of the seismicity rate changes is analyzed by generating LTA(t) function at the nodes of 0.05ºx0.05º grid spaces. Using a moving time window T =5.5 years, seismic W

quiescence Z-value distribution in the beginning of 2014 is mapped. The regions exhibiting lower b-values and higher Z-values are observed in the Burdur fault zone, Aliağa-Dumlu-pınar faults and Bakırçay grabens. As a general result, the re-lationships between seismic b-value, fractal dimension Dc-value and seismic quiescence Z-value may provide significant clues to reveal the probable locations of future earthquakes in the Western Anatolian region.

AcknowledgementsThe author would like to thank to Prof. Dr. Stefan Weimer

for providing ZMAP software and the reviewers, especially Ewald Brückl, for their useful and constructive suggestions in improving this paper. I am grateful to Dr. Dogan Kalafat (KOERI) for providing us the earthquake catalogue. I also thank to KOERI for providing earthquake database via internet. This study is supported by Gümüşhane University Scientific Re-search Project (GUBAP, Turkey) with project no 13.F5117.02.1.

________________________

ReferencesAki, K., 1965.

N=a-bM and its confidence limits. Bulletin of the Earthquake Research Institute, Tokyo University, 43, 237-239.

Maximum likelihood estimate of b in the formula

Aki, K., 1981.pp 566-574.

Arabasz, W.J. and Hill, S.J., 1996. Applying Reasenberg’s clus-ter analysis algorithm to regional earthquake catalogs out-side California (abstract). Seismological Research Letters, 67 (2), p. 30.

Barton, D.J., Foulger, G.R., Henderson, J.R. and Julian, B.R., 1999. Frequency-magnitude statistics and spatial correlation di-mensions of earthquakes at Long Valley Caldera, California. Geophysical Journal International, 138, 563–570. http://dx. doi.org/10.1046/j.1365-246X.1999.00898.x

Bayrak, Y., Öztürk, S., Çınar, H., Kalafat, D., Tsapanos, T.M., Kora-vos, G.Ch. and Leventakis, G.A., 2009. Estimating earthquake hazard parameters from instrumental data for different re-gions in and around Turkey. Engineering Geology, 105, 200-210. http://dx.doi.org/10.1016/j.enggeo.2009.02.004

Bozkurt, E., 2001. Neotectonics of Turkey-a synthesis. Geodi-namica Acta, 14, 3-30. http://dx.doi.org/10.1016/S0985-3111(01)01066-X

Carrol, R.J. and Ruppert, D., 1996. The use and misuse of ortho-gonal regression estimation in linear errors-in-variables mo-dels. The American Statistician, 50, 1-6.

Erdik, M., Alpay, B.Y., Onur, T., Sesetyan, K. and Birgoren, G., 1999. Assessment of earthquake hazard in Turkey and neigh-boring regions. Annali di Geofisica, 42, 1125-1138.

Frohlich, C. and Davis, S., 1993. Teleseismic b-values: Or, much ado about 1.0. Journal of Geophysical Research, 98 (B1), 631-644. http://dx.doi.org/10.1029/92JB01891

Grassberger, P. and Procaccia, I., 1983. Measuring the strange-ness of strange attractors, Physica, 9(D), 189–208.

Gutenberg, R. and Richter, C.F., 1944. Frequency of earthqua-kes in California. Bulletin of the Seismological Society of America, 34, 185-188.

Hirata, T., 1989. Correlation between the b-value and the frac-tal dimension of earthquakes. Journal of Geophysical Re-search. 94, 7507-7514. http://dx.doi.org/10.1029/JB094iB06 p07507.

İlkışık, O.M., 1995. Regional heat flow in western Anatolia us-ing silica temperature estimates from thermal springs. Tec-tonophysics, 244, 175-184. http://dx.doi.org/10.1016/0040-1951(94)00226-Y

Joswig, M., 2001. Mapping seismic quiescence in California. Bulletin of the Seismological Society of America, 91/1, 64-81. http://dx.doi.org/10.1785/0119980160

Kagan, Y.Y., 2007. Earthquake spatial distribution: the correla-tion dimension. Geophysical Journal International, 168, 1175-1194. http://dx.doi.org/10.1111/j.1365-246X.2006.03251.x

Lopez Casado, C., Sanz de Galdano, C., Delgado, J. and Peina-do, M.A., 1995. The b parameter in the Betic Cordillera, Rif and nearbysectors. Relations with the tectonics of the region. Tectonophysics, 248, 277–292. http://dx.doi.org/10.1016/0040-1951(94)00278-H.

McClusky, S., Balassanian, S., Barka, A., Demir, C., Ergintav, S., Georgiev, I., Gürkan, O., Hamburger, M., Kahle, K.H.H., Kastens, K., Kekelidze, G., King, R., Kotzev, V., Lenk, O., Mahmoud, S.,

Earthquake prediction. AGU, vol 4. Washington,

182

Serkan ÖZTÜRK

K., Kekelidze, G., King, R., Kotzev, V., Lenk, O., Mahmoud, S., Mishin, A., Nadariya, M., Ouzounis, A., Paradissis, D., Peter, Y., Prilepin, M., Reilinger, R., Şanlı, I., Seeger, H., Tealeb, A., Toksöz, M.N. and Veis, G., 2000. Global positioning system constraints on plate kinematics and dynamics in the eastern Mediterra-nean and Caucasus. Journal of Geophysical Research, 105 (B3), 5695-5719. http://dx.doi.org/10.1029/1996JB900351

Mandelbrot, B.B., 1982. The fractal geometry of nature. Free-man Press, San Francisco.

Oral, M.,B., Reilinger, R.,E., Toksöz, M.,N., Kong, R., W., Barka, A.,A., Kınık, I. and Lenk, O., 1995. Global positioning system offers evidence of plate motions in eastern Mediterranean. EOS, Transactions American Geophysical Union, 76, 9. http:// dx.doi.org/10.1029/EO076i002p00009-01

Öncel, A.O., Alptekin, Ö. and Main, I.G., 1995. Temporal varia-tions of the fractal properties of seismicity in the western part of the North Anatolian fault zone: possible artifacts due to improvements in station coverage. Nonlinear Processes in Geophysiscs. 2, 147-157. http://dx.doi.org/10.5194/npg-2-147-1995

Öncel, A.O., Main, I.G., Alptekin, Ö. and Cowie, P.A., 1996. Tem-poral variations of the fractal properties of seismicity in the north Anatolian fault zone between 31°E and 41°E. Pure and Applied Geophysics, 146, 148-159. http://dx.doi.org/10.1007/ BF00876441

Öncel, A.O., Wilson, T.H. and Nishizawa, O., 2001. Size scaling relations in the active fault networks of Japan and their cor-relation with Gutenberg–Richter b-Values. Journal of Geo-physical Research, 106/B10, 21827–21841. http://dx.doi.org/ 10.1029/2000JB900408.

Öncel, A.O. and Wilson, T.H., 2002. Space-time correlations of seismotectonic parameters and examples from Japan and Turkey preceding the İzmit earthquake. Bulletin of the Seis-mological Society of America, 92, 339-350. http://dx.doi.org/ 10.1785/0120000844

Öncel A.O. and Wilson, T.H., 2004. Correlation of seismotec-tonic variables and GPS strain- measurements in western Turkey. Journal of Geophysical Research, 109/B11, B11306. http://dx.doi.org/10.1029/2004JB003101

Öncel, A.O. and Wilson, T.H., 2007. Anomalous seismicity pre-ceding the 1999 Izmit event, NW Turkey. Geophysical Jour-nal International, 169, 259-270. http://dx.doi.org/10.1111/ j.1365-246X.2006.03298.x.

Öztürk, S., Y. Bayrak, H. Çinar, G.Ch. Koravos and T.M. Tsapa-nos, 2008. A quantitative appraisal of earthquake hazard parameters computed from Gumbel I method for different regions in and around Turkey, Natural Hazards 47, 471-495. http://dx.doi.org/10.1007/s11069-008-9234-6

Öztürk, S., 2009. An application of the earthquake hazard and aftershock probability evaluation methods to Turkey earth-quakes. PhD Thesis, Karadeniz Technical University, Trabzon, 346 pp.

Öztürk, S., 2011. Characteristics of Seismic Activity in the Wes-tern, Central and Eastern Parts of the North Anatolian Fault Zone, Turkey: Temporal and Spatial Analysis. Acta Geophy-

sica, 59/2, 209-238. http://dx.doi.org/10.2478/s11600-010-0050-5

Öztürk, S., 2012. Statistical correlation between -value and fractal dimension regarding Turkish epicentre distribution. Earth Sciences Research Journal, 16/2, 103,108.

Öztürk, S. and Bayrak, Y., 2012. Spatial variations of precursory seismic quiescence observed in recent years in the eastern part of Turkey. Acta Geophysica, 60/1, 92-118. http://dx.doi. org/10.2478/s11600-011-0035-z

Öztürk, S., 2013. A statistical assessment of current seismic quiescence along the North Anatolian Fault Zone: Earth-quake precursors. Austrian Journal of Earth Sciences, 106/2, 4-17.

Polat, O., Gok, E. and Yılmaz, D., 2008. Earthquake hazard of the Aegean Extension region (West Turkey). Turkish Journal of Earth Sciences, 17, 593-614.

Reasenberg, P.A., 1985. Second-order moment of Central Cali-fornia seismicity, 1969-1982. Journal of Geophysical Research, 90/B7, 5479-5495. http://dx.doi.org/10.1029/JB090iB07p05479

Richter, C.F., 1935. An instrumental magnitude scale, Bulletin of the Seismological Society of America, 25, 1-32.

Richter, C.F., 1938. Elementary seismology, W.H. Freeman and Co., San Francisco, California, 758 pp., Bulletin of the Seismo-logical Society of America, 25, 1-32.

Roy, S., Ghosh, U., Hazra, S. and Kayal, J.R., 2011. Fractal dimen-sion and b-value mapping in the Andaman-Sumatra subduc-tion zone. Natural Hazards, 57, 27–37. http://dx.doi.org/10. 1007/s11069-010-9667-6

Sayıl, N. and Osmanşahin, I., 2008. An investigation of seismi-city for western Anatolia. Natural Hazards, 44, 51–64. http:// dx.doi.org/10.1007/s11069-007-9141-2

Scholz, C.H., 1968. The frequency-magnitude relation of micro-fracturing in rock and its relation to earthquakes. Bulletin of the Seismological Society of America, 58, 399-415.

Schorlemmer, D., Wiemer S., and Wyss, M., 2005. Variations in earthquake-size distribution across different stress regimes. Nature, 437, 539-542. http://dx.doi.org/10.1038/nature04094.

Şaroğlu, F., Emre, O. and Kuşcu, O., 1992. Active fault map of Turkey, General Directorate of Mineral Research and Explo-ration, Ankara, Turkey.

Utsu, T., 1971. Aftershock and earthquake statistic (III): Analy-ses of the distribution of earthquakes in magnitude, time and space with special consideration to clustering charac-teristics of earthquake occurrence (1). Journal of Faculty of Science, Hokkaido University, Series VII (Geophysics), 3, 379-441.

Westerhaus, M., Wyss, M., Yılmaz, R. and Zschau, J., 2002. Cor-relating variations of b values and crustal deformations du-ring the 1990s may have pinpointed the rupture initiation of the Mw= 7.4 İzmit earthquake of 1999 August 17. Geophy-sical Journal International, 148, 139–152. http://dx.doi.org/ 10.1046/j.0956-540x.2001.01554.x.

Wiemer, S., and Wyss, M., 1994. Seismic quiescence before the Landers (M=7.5) and Big Bear (6.5) 1992 earthquakes. Bulle-tin of the Seismological Society of America, 84/3, 900-916.

b

_

183


Wiemer, S. and Wyss, M., 2000.pleteness in earthquake catalogs: Examples from Alaska, the Western United States, and Japan. Bulletin of the Seismolo-gical Society of America, 90 (3), 859-869. http://dx.doi.org/ 10.1785/0119990114

Wiemer, S., 2001. A software package to analyze seismicity: ZMAP. Seismological Research Letters, 72/2, 373-382. http:// dx.doi.org/10.1785/gssrl.72.3.373

Woessner, J. and Wiemer, S., 2005. Assessing the quality of earthquake catalogues: Estimating the magnitude of com-pleteness and its uncertainty. Bulletin of the Seismological Society of America, 95/2, 684-698. http://dx.doi.org/10.1785/ 0120040007

Wyss, M., Sobolev, G.A. and Clippard, J.D., 2004. Seismic quie-scence precursors to two M7 earthquakes on Sakhalin Island, measured by two methods. Earth Planets Space, 56, 725-740.

Minimum magnitude of com-

Received: 02 May 2014Accepted: 18 August 2015

Serkan ÖZTÜRKGümüşhane University, Department of Geophysics, TR-29100, Gümü-

şhane, Turkey;

[email protected]

184

a study on the correlations between seismotectonic -value ... · regions. in addition, a few new...

Documents